The significant increase in complexity of Exascale platforms due to energy-constrained, billion-way parallelism, with major changes to processor and memory architecture, requires new energy-efficient and resilient programming techniques that are portable across multiple future generations of machines. We believe that guaranteeing adequate scalability, programmability, performance portability, resilience, and energy efficiency requires a fundamentally new approach, combined with a transition path for existing scientific applications, to fully explore the rewards of todays and tomorrows systems. We present HPX -a parallel runtime system which extends the C++11/14 standard to facilitate distributed operations, enable fine-grained constraint based parallelism, and support runtime adaptive resource management. This provides a widely accepted API enabling programmability, composability and performance portability of user applications. By employing a global address space, we seamlessly augment the standard to apply to a distributed case. We present HPX's architecture, design decisions, and results selected from a diverse set of application runs showing superior performance, scalability, and efficiency over conventional practice.

The scalability and efficiency of graph applications are significantly constrained by conventional systems and their supporting programming models. Technology trends like multicore, manycore, and heterogeneous system architectures are introducing further challenges and possibilities for emerging application domains such as graph applications. This paper explores the space of effective parallel execution of ephemeral graphs that are dynamically generated using the Barnes-Hut algorithm to exemplify dynamic workloads. The workloads are expressed using the semantics of an Exascale computing execution model called ParalleX. For comparison, results using conventional execution model semantics are also presented. We find improved load balancing during runtime and automatic parallelism discovery improving efficiency using the advanced semantics for Exascale computing. I. INTRODUCTIONA large class of problems in physics and molecular biology can be represented using a particle interaction method commonly known as N-Body and computational techniques based on these discretization methods. The science domains utilizing the particle interaction discretization model are limited by the number of particles that can be simulated and the time it takes to execute the computational techniques. Conventional practices have significantly advanced particle interaction based methodologies. However, the combined ecosystem of emerging multicore based system architectures and conventional programming models are imposing grave challenges to the continued effectiveness of these methods. This research identifies and addresses these challenges through the hypothesis that emerging system architectures and extreme scale oriented runtime systems can dramatically improve the end science.Applications based on graphs and tree data structures rely on more dynamic, adaptive, and irregular computations. This work explores an exemplar dynamic tree based application embodied by an N-Body simulation. Systems comprising many particles (N-Body problem) interacting through long-range forces have considerable computational science interest. N-Body systems comprising three or more particles do not have a closed form solution; consequently, iterative methods are used to approximate solutions for the N-Body problem. The N-Body problem simulates the evolution of n particles under the influence of mutual pairwise interactions through forces such as gravitational pull or electrostatic forces. This work focuses on gravitational forces operating on the N-Body system and the Barnes-Hut approximation of the N-Body solution.While several approaches to simulating N-Body systems exist, the Barnes-Hut algorithm [5] is widely used in astrophysical simulations mainly due to its logarithmic computational complexity while generating results that are within acceptable bounds of accuracy. In the Barnes-Hut algorithm the particles are grouped by a hierarchy of cube structures using a recursive algorithm which subdivides the cubes until there is one particle per sub-cube. It then u...

We test a new "hybrid" scheme for simulating dynamical fluid flows in which cylindrical components of the momentum are advected across a rotating Cartesian coordinate mesh. This hybrid scheme allows us to conserve angular momentum to machine precision while capitalizing on the advantages offered by a Cartesian mesh, such as a straightforward implementation of mesh refinement. Our test focuses on measuring the real and imaginary parts of the eigenfrequency of unstable axisymmetric modes that naturally arise in massless polytropic tori having a range of different aspect ratios, and quantifying the uncertainty in these measurements. Our measured eigenfrequencies show good agreement with the results obtained from the linear stability analysis of Kojima (1986) and from nonlinear hydrodynamic simulations performed on a cylindrical coordinate mesh by Woodward et al. (1994). When compared against results conducted with a traditional Cartesian advection scheme, the hybrid scheme achieves qualitative convergence at the same or, in some cases, much lower grid resolutions and conserves angular momentum to a much higher degree of precision. As a result, this hybrid scheme is much better suited for simulating astrophysical fluid flows, such as accretion disks and mass-transferring binary systems.

No abstract

No abstract